Relationship between DNA Adduct Levels, Repair ... - CiteSeerX

Vol. 10, 749 –758, November 1999

Cell Growth & Differentiation

Relationship between DNA Adduct Levels, Repair Enzyme, and Apoptosis as a Function of DNA Methylation by Azoxymethane1 Mee Young Hong, Robert S. Chapkin, Christopher P. Wild, Jeffrey S. Morris, Naisyin Wang, Raymond J. Carroll, Nancy D. Turner, and Joanne R. Lupton2 Faculty of Nutrition [M. Y. H., R. S. C., R. J. C., N. D. T., J. R. L.], and the Department of Statistics [J. S. M., N. W., R. J. C.], Texas A & M University, College Station, Texas 77843-2471, and Molecular Epidemiology Unit, School of Medicine, University of Leeds, Leeds, LS2 9JT United Kingdom [C. P. W.]

Abstract DNA alkylating agent exposure results in the formation of a number of DNA adducts, with O6-methyldeoxyguanosine (O6-medG) being the major mutagenic and cytotoxic DNA lesion. Critical to the prevention of colon cancer is the removal of O6-medG DNA adducts, either through repair, for example, by O6-alkylguanineDNA alkyltransferase (ATase) or targeted apoptosis. We report how rat colonocytes respond to administration of azoxymethane (a well-characterized experimental colon carcinogen and DNA-methylating agent) in terms of O6-medG DNA adduct formation and adduct removal by ATase and apoptosis. Our results are: (a) DNA damage is greater in actively proliferating cells than in the differentiated cell compartment; (b) expression of the DNA repair enzyme ATase was not targeted to the proliferating cells or stem cells but rather is confined primarily to the upper portion of the crypt; (c) apoptosis is primarily targeted to the stem cell and proliferative compartments; and (d) the increase in DNA repair enzyme expression over time in the bottom one-third of the crypt corresponds with the decrease in apoptosis in this same crypt region.

Introduction Colon tumors develop from a series of somatic mutations subsequent to an initiating event such as DNA methylation. Experimental carcinogens known to produce gastrointestinal tumors in rodents (1,2-dimethylhydrazine and its derivative

Received 6/2/99; revised 8/25/99; accepted 9/14/99. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 This research was supported by Grants CA 61750 (to J. R. L.), CA59034 (to R. S. C.), CA57030 (to R. J. C.), CA74552 (to N. W.), and National Institute of Environmental Health Science Grant P30-ES09106 from the NIH. 2 To whom requests for reprints should be addressed, at Faculty of Nutrition, 218 Kleberg Building, Texas A & M University, College Station, TX 77843-2471. Phone: (409) 845-2142; Fax: (409) 862-2378; E-mail: [email protected].

azoxymethane) are alkylating agents (1, 2), and alkylating agents are also documented causes of cancer in humans as a result of lifetime environmental exposure (3, 4). DNA alkylating agents result in the formation of a number of DNA adducts, with O6-medG3 being the major mutagenic and cytotoxic DNA lesion resulting from methylating agents (5–7). During DNA replication, these O6-medG DNA adducts induce G-to-A transitions (8, 9), a common mutation reported in the activation of oncogenes and inactivation of tumor suppressor genes within tumors induced by alkylating agents (3, 10). Critical to the prevention of colon cancer is the removal of O6-medG DNA adducts, either through repair, for example, by ATase (EC 2.1.1.63) or targeted apoptosis. ATase is an inducible repair enzyme (7) that acts by transferring the methyl group from guanine in DNA to a cysteine residue on the repair protein (11), resulting in rapid removal of the promutagenic O6-medG DNA adducts and irreversible inactivation of the repair enzyme (7). ATase has been shown to operate similarly in humans and rats (12). In addition to this direct repair of DNA damage, lesions can be eliminated by nucleotide excision. For a recent review, see Sancar (13). Cells with unrepaired DNA adducts may be eliminated through the activation of the apoptotic cascade, resulting in their selective removal (14). During the initiation of tumorigenesis after carcinogen treatment, there is an immediate apoptotic response to DNA damage in the colonic epithelium (15). In fact, Meikrantz et al. (16) have shown that O6-medG DNA lesions may trigger apoptosis. Although the effect of DNA alkylating agent administration on O6-medG DNA adduct formation (17, 18), DNA repair (19), and apoptosis (15) have been reported independently, there is no information on the interrelationship among the three. We hereby report how rat colonocytes respond to administration of azoxymethane (a well-characterized experimental colon carcinogen and DNA alkylating agent) in terms of O6medG DNA adduct formation and adduct removal by ATase and apoptosis. In addition, we show how this relationship changes during the first 12 h of the initiation process and the specific localization within the colonic crypt of adducts, repair enzyme, and apoptotic cells.

Results and Discussion The localization of DNA damage, repair, and targeted apoptosis within the crypt is important because there is a distinct

3 The abbreviations used are: O6-medG, O6-methyldeoxyguanosine; ATase, O6-alkylguanine-DNA alkyltransferase; AOM, azoxymethane; DMH, 1,2-dimethylhydrazine; TdT, 59-terminal deoxynucleotidyl transferase.

749

750

DNA Damage, Repair, and Apoptotic Removal

Fig. 1. DNA adduct levels in colonocytes over time after AOM injection. DNA adduct levels increased throughout the 12-h time course. Points not having the same superscript letters are significantly different at P , 0.05. Bars, SE.

hierarchical arrangement of colonocytes that has been welldocumented (20). Stem cells are located toward the base of the crypt, actively proliferating cells occupy the lower twothirds of the crypt, and the upper one-third of the crypt contains fully differentiated cells. Spontaneous apoptosis generally occurs toward the lumenal surface, and cells are then exfoliated into the fecal stream. Thus, if one knows the position of a cell within the crypt, assumptions can be made about the stage of its life cycle. In this study, the localization of DNA damage, repair, and targeted apoptosis within the colonic crypt were detected using computerized quantitative immunohistochemistry with digital image analysis. This powerful technique allowed for the simultaneous analysis of all three parameters in situ in colons of the same rats. We report four significant findings. DNA Damage Is Greater in Actively Proliferating Cells Than in the Differentiated Cell Compartment. DNA adduct levels increased throughout the 12-h time course (P , 0.05; Fig. 1). The absolute level of O6-methylguanine was not determined in this study, but a previous report in male Fischer 344 rats with two injections of 15 mg/kg body weight of AOM, separated by 1 week, reported adduct levels of 232 6 24 mmol O6-methylguanine/mol guanine in DNA extracted from whole colon (21). Fig. 2 shows a typical colonic crypt assayed for DNA adducts at time zero (Fig. 2A) and 9 h after carcinogen injection (Fig. 2B). Adducts are barely detectable at time zero, but colon cells are intensely stained for DNA adducts at 9 h. Although adduct formation increased with time, the distribution of adducts within the crypt (from the base of the crypt at designated position 0.0 to the top of the crypt at designated position 1.0) remained the same throughout the time course of the experiment (Fig. 3). DNA adduct levels were slightly but significantly higher in the bottom one-third of the crypt (note the upward slope of the line toward the lower relative cell position in Fig. 3). This significantly higher level of DNA adducts in the bottom onethird of the crypt is summarized for all times in Fig. 4. DNA damage as assessed by immunohistochemistry is the net effect of damage and repair. Thus, the greater damage seen in the actively proliferating compartment could be attributable to a greater rate of damage and/or a lower rate of

repair. It is unlikely that there was a greater rate of damage in these proliferating cells, because alkylation (unlike other forms of DNA damage that may require intercalation of the agent into DNA) does not appear to be cell cycle specific (22). In contrast, a number of studies have shown cell cycle specificity with respect to DNA repair. For example, Smith et al. (23) measured the removal of selected alkylation products from DNA of synchronized 10T1⁄2 cells. Their study clearly showed that O6-methylguanine residues were excised from DNA during the G1 phase of the cell cycle but not during the S phase (23). In addition, our study shows localization of DNA repair enzyme predominantly in the nonproliferative compartment. It appears that lack of repair rather than higher rates of DNA damage in proliferating cells accounts for the slightly greater amount of DNA damage in the proliferative compartment. Although the magnitude of the differences in DNA damage amounts was not great, damage in this compartment is biologically important because DNA damage can be converted to mutations during cell division, an important step in the tumorigenic process. DNA Repair Enzyme Does Not Target the Proliferating Cells or Stem Cells. Expression of the DNA repair enzyme increased between 6 and 9 h and then appeared to plateau through 12 h (Fig. 5). There is a reported strong correlation between quantitative immunohistochemistry and biochemical alkyltransferase activity (P , 0.0001; Ref. 24), suggesting that our measurements were measuring active DNA repair. It should be noted, however, that our quantitative immunohistochemical measurements were designed to detect the localization of the repair enzyme and changes in its expression over time, not the actual removal of DNA adducts. We do not know the level of enzyme expression at which DNA adduct repair might become limiting. There was a point of inflection in DNA repair enzyme staining intensity at 6 h; however, the overall curve for DNA repair was not significantly different over time. At every time point except for 12 h, DNA repair enzyme was much higher in the upper portion of the crypt (toward 1.0) than it was toward the base of the crypt (position 0.0; Fig. 6). Of interest is the finding that the pattern of DNA repair enzyme expression changed at 12 h in that expression in the bottom of the crypt significantly increased as compared with previous time points (P , 0.05; Fig. 6). It is possible that this increased expression may represent both an increase in protein expression (via transcriptional induction; active protein) and also an accumulation of inactive protein targeted for degradation, a distinction that would have to be resolved by measuring enzyme activity. The purpose of the present study was to detect the localization of the repair enzyme, which is not possible, when using scraped mucosa for enzyme assays. An analysis of DNA repair enzyme levels over the entire course of the 12-h experiment shows that expression is highest in the top one-third of the crypt, intermediate in the middle one-third, and lowest in the bottom one-third (Fig. 7). Of potential biological significance is that the lower expression of DNA repair enzyme level in the bottom one-third of the crypt correlates with the higher levels of DNA adducts in this crypt portion (P , 0.05). The only other study documenting the localization of DNA damage repair enzyme in the colonic crypt reported similar


Fig. 2. Photomicrograph (340) of colonic crypts stained for O6-methylguanine adduct formation. A, 0 h, no carcinogen injection. B, 9 h after carcinogen injection. At time 0, little staining is discernible. In comparison, at 9 h intense staining is observed.

Fig. 3. DNA adduct levels as a function of relative crypt cell position. Position 0.0 indicates the base of the crypt; 1.0 designates cells at the lumenal surface. There is a relatively uniform distribution of label throughout the crypt, with a slight, but significantly (P , 0.05) higher level of expression toward the base.

findings (24), i.e., a gradient of ATase expression, with higher levels detected toward the top of the crypt. These authors attributed the higher expression of ATase toward the lumenal surface as a necessary protective mechanism because the surface cells are likely to be the first targets for ingested or endogenously produced carcinogens. AOM is known to be

metabolized to the ultimate carcinogen, primarily by the colonic microflora, although some metabolism of the carcinogen can also occur in the colonocytes themselves (25). Thus, the upper crypt cells would have more direct exposure to the carcinogen. It is also possible, although it has not been reported, that there are variations in carcinogen metabolism

751

752


Fig. 4. DNA adduct levels as a function of cell location (tertile) within the colonic crypt. When all data over time are combined, DNA adduct levels are higher in the lower one-third of the crypt (stem cell region) compared with the middle or upper portion of the crypt. Columns not having the same superscript letters are significantly different at P , 0.05. Bars, SE.

in the different cells that populate the crypt attributable to differences in P-450 expression. Apoptosis Is Primarily Targeted to the Stem Cell and Proliferative Compartments. There was a highly significant increase (P , 0.001) in apoptosis after carcinogen injection (Fig. 8). Maximum apoptosis was achieved by 9 h (an apoptotic index of 18.8%) and decreased significantly (P , 0.05) between 9 and 12 h (Fig. 8). In noncarcinogen-injected animals, apoptotic cells averaged ,1 per crypt and were primarily located at the top of the crypt, as shown in Fig. 9A. However, after carcinogen injection, most apoptotic cells were found in the bottom one-third of the crypt, where the stem cell population is located (Fig. 9B). This shift in localization of apoptotic cells in response to carcinogen injection was highly significant (P , 0.001). The distribution of apoptotic cells within the colonic crypt is illustrated in Fig. 10, with 0.0 being the bottom of the crypt and 1.0 representing the lumenal surface. From the time that apoptosis is discernible above baseline (3 h), it is always greater at the crypt base than at the top (Fig. 10). This is clearly demonstrated when all time points are combined (Fig. 11), indicating that the apoptotic index is greatest in the bottom one-third of the crypt, lower in the middle one-third, and still lower in the upper one-third (P , 0.05). Ijiri and Potten (26), in an important series of experiments, tested the effect of 18 different cytotoxic agents on the spatial distributions of cell death in mouse small intestinal crypts at various times from 3 to 12 h after administration of each agent. Interestingly, for the five alkylating agents used, the peak of apoptotic activity was seen at 9 h after administration (27), which is consistent with our data for the alkylating agent AOM. In a follow-up study from this group, a comparison was made between the small and large intestine in the time point for the highest apoptotic response (20). The colon carcinogen DMH, a precursor to AOM, showed a peak apoptotic response in the colonic crypt at 8 h after injection (apoptosis was not measured at 9 h). Thus, our finding of a peak response at 9 h and a drop off after that time is con-

Fig. 5. DNA repair enzyme (ATase) levels in colonocytes as a function of time after AOM injection. Although there was a numerical increase in DNA repair enzyme staining intensity between 6 and 9 h, no significant difference in repair enzyme expression over time (P . 0.05) was detected. Bars, SE.

sistent with the literature. Also, in our study the peak apoptotic response was an apoptotic index of 18.8%, which compares favorably with Potten’s peak apoptotic response using DMH of 19.5% (20). In contrast to the work of Potten’s group (20), which indicated apoptosis was confined to cells above the purported stem cell region, we found a large proportion of the apoptotic cells throughout the bottom one-third of the crypt. Potten et al. (20) argue that a possible reason for the much higher incidence of carcinoma in the large intestine as compared with the small intestine is that apoptosis targets the stem cell population in the small intestine but not in the large intestine. The absence of selective deletion in the large intestine might result in the stability and clonal expansion of DNA-damaged cells. Our data suggest that apoptosis targets the entire proliferative compartment equally, including stem cells and other actively proliferating cells. Furthermore, we, unlike Potten, measured DNA damage in the same rats and found that it was not specifically targeted to the stem cell or proliferative compartment but rather, evenly distributed throughout the entire crypt. Thus, it does not appear that the much higher incidence of cancer in the large bowel rather than the small bowel can be accounted for by apoptosis targeting DNA damage in the small intestine but not in the colon. Although cell division was not measured in the present study, because division is necessary to convert adducts to mutations (28), it will certainly play an important role in addition to apoptosis in determining future tumor development. In documenting the distribution of apoptotic cells within the crypt throughout the time course of the study, it was expected that we would observe a shift of peak apoptotic incidence to higher cell positions over time as cells toward the bottom of the crypt engulfed apoptotic bodies and migrated up the crypt. No shift to the right was observed. The reason for this lack of shift in peak apoptotic activity is not known but could involve a decrease in migration rate or an extrusion of apoptotic bodies into the crypt lumen rather than engulfment by neighboring cells and migration to the lumenal surface. Interestingly, Potten et al. (20) found the same lack


Fig. 6. DNA repair enzyme (ATase) levels as a function of relative cell position within the colonic crypt. On the X axis, 0.0 is the base of the crypt, and 1.0 is the lumenal surface of the crypt. From 0 to 9 h, expression of the repair enzyme is highest in the upper portion of the crypt. At 12 h, expression in the lower portion of the crypt increased significantly as compared with previous time points (P , 0.05).

Fig. 7. DNA repair enzyme levels are highest in the upper portion of the crypt, intermediate in the middle, and lowest in the lower one-third of the crypt. Columns not having the same superscript letters are significantly different at P , 0.05. Bars, SE.

Fig. 8. Apoptotic index in colonocytes as a function of time after AOM injection. The apoptotic index is the proportion of cells in the colonic crypt column undergoing apoptosis. Maximum apoptosis was achieved by 9 h after carcinogen injection. Points not having the same superscript letters are significantly different at P , 0.001. Bars, SE.

of shift with DMH in his experiments as we found with AOM in ours. Our in vivo finding that apoptosis is highest in the region of lowest expression of DNA repair enzyme is consistent with

the findings of a recent in vitro study (16). Meikrantz et al. (16), using isogenic Chinese hamster ovary cell lines that stably express the human ATase repair enzyme, examined the ability of the cells with high levels of repair enzyme to

753

754


Fig. 9. Photomicrograph (340) of apoptotic cells. A, 0 h, no carcinogen injection. B, 9 h after carcinogen injection. The box outlines apoptotic cells. At time zero, few apoptotic cells are evident, and most are located toward the top of the crypt. In contrast, 9 h after carcinogen injection, the number of apoptotic cells increases significantly, and most are located toward the base of the crypt.

Fig. 10. Apoptotic index as a function of relative cell position within the colonic crypt. In each panel, 0.0 is the bottom of the crypt, whereas 1.0 is at the lumenal surface. As shown in these figures, most apoptotic cells are found toward the base of the crypt after carcinogen injection.

undergo methylation-induced apoptosis compared with cells not expressing ATase. In response to treatment with the methylating agent N-methyl-N9-nitro-N-nitrosoguanidine, the cells expressing the repair enzyme underwent limited apoptosis compared with vector controls, which exhibited only 0.5% survival (16). A second study from the same laboratory further investigated the signaling pathway from DNA

damage to apoptotic response (29). In this case, they used isogenic mouse cell lines either proficient or deficient in another DNA alkylation repair enzyme, a DNA glycosylase that repairs 3-methyladenine. This lesion leads to stalled replication, induction of strand breaks, and apoptosis. In contrast, the major apoptotic pathway with the O6-medG lesion is via mismatch repair. However, this is only relevant


Fig. 11. Apoptotic index within the colonic crypt. The index is highest in the lower portion of the crypt, lower in the middle portion, and still lower in the uppermost one-third of the crypt. Columns not having the same superscript letters are significantly different at P , 0.05. Bars, SE.

when ATase activity is deficient. Lesions then persist through a second round of DNA replication, and the mismatch repair activity results in strand breaks and apoptosis. This has been effectively demonstrated in a mouse model with targeted inactivation of the Msh2 mismatch repair gene and using O6-benzylguanine to inactivate ATase (30). The Increase in Repair Enzyme Expression Coincides with the Decrease in Apoptosis. Fig. 12 shows the relationship between DNA adduct level, DNA repair enzyme, and apoptosis over time. DNA adduct level continues to increase with time. The DNA repair enzyme appears to rise after 6 h and plateaus at 9 h. Apoptosis increases over time, peaking at 9 h, and then decreases between 9 and 12 h. The distribution of DNA adduct levels, DNA repair enzyme, and apoptosis within the crypt at each time point is shown in Fig. 13. The distribution of adducts (skewed toward the bottom of the crypt) remains the same throughout the 12-h time period. At every time point other than 12 h, DNA repair enzyme expression is highest toward the top of the crypt, and at all time points, other than 0 h, apoptosis is targeted toward the bottom one-third of the crypt (Fig. 13). Of interest is our finding that the localization of repair enzyme changes dramatically at 12 h in that it becomes uniformly distributed throughout the crypt. The reason for this change in repair enzyme localization is not known, but the timing of increased repair enzyme expression in the lower part of the crypt coincides with a decrease in apoptosis in this same region. One might speculate that as apoptotic removal of DNA damaged cells in the lower portion of the crypt is diminished, DNA repair is enhanced as a compensatory mechanism. Alternatively, if DNA damage elicits the apoptotic response, (16), then with less DNA damage (because of greater repair) less apoptosis should be elicited. It should be noted, however, that this hypothesized relationship between ATase expression and apoptosis is not supported by an extensive literature (13). In summary, we report, for the first time, the relationship between DNA damage, repair, and apoptosis as a function of position within the colonic crypt within 12 h of administration

of the alkylating agent AOM. There were four significant observations: (a) DNA damage is greater in actively proliferating cells than in the differentiated cell compartment. This is likely because of a lower rate of repair in the proliferative compartment rather than a greater rate of damage; (b) DNA repair enzyme does not target the proliferating cells or stem cells but rather is confined primarily to the upper portion of the crypt. The important point here is that most studies assume that the DNA repair enzyme targets DNA damaged cells. It appears from our study that the enzyme does not specifically target DNA-damaged cells but rather is localized to cells near the lumenal surface prior to 12 h after AOM injection; (c) apoptosis is found predominantly in the region of actively cycling cells, where repair enzyme expression is lowest; and (d) the increase in DNA repair enzyme expression in the bottom one-third of the crypt corresponds with the timing of the decrease in apoptosis in this same crypt region. This may indicate an interactive compensatory response.

Materials and Methods Animals. Male weanling Sprague Dawley rats (Harlan Sprague Dawley, Houston, TX) were used after approval of our protocol by the University Laboratory Animal Care Committee of Texas A & M University. Animal care and handling conformed to NIH guidelines. Rats were individually housed and maintained in a temperature- and humidity-controlled animal facility with a daily 15:9-h light:dark photoperiod. A defined diet and water were freely available. Specifics of the diet have been reported in detail (31). Administration of Carcinogen. AOM (Sigma Chemical Co., St. Louis, MO) was injected s.c. (15 mg/kg) precisely at 9 a.m. to minimize the known diurnal variations in apoptosis (32). Each animal was terminated 3, 6, 9, or 12 h after injection. Time zero represents a negative control, because no injection was performed. Three rats were used for each time point. Tissue Acquisition for Measurements of DNA Adduct Levels, Repair Enzyme, and Apoptosis. Rats were euthanized by CO2 gas, and the entire colon was immediately resected. After removal of the rectum, the last 2 cm of the distal colon were taken for immunohistochemistry and divided in half lengthwise. One-half of the distal colon was fixed in 70% ethanol overnight, and the other half was fixed in a 4% paraformaldehyde solution for 4 h, followed by washing with 50 and 70% ethanol. All segments were placed in graded levels of ethanol, from 80 to 100%, and finally into xylene before they were embedded in paraffin wax. Ethanolfixed paraffin sections (4 mm thick) were cut perpendicular to the mucosal surface and affixed to poly-L-lysine slides for measurement of O6-medG DNA adduct levels. Paraformaldehyde-fixed, paraffin-embedded tissue sections (4 mm thick) were placed onto Superfrost Plus slides and stored at 280°C until use for the apoptosis and repair enzyme assays. In Vivo Measurements of DNA Adducts, Apoptosis, and Repair Enzyme. For in vivo measurement of O6-medG, ethanol-fixed, paraffinembedded tissue sections were used for measuring O6-medG DNA adduct levels as described previously (33). Tissues on slides were deparaffinized and rehydrated. Tissue sections were incubated with 0.3% H2O2 in methanol to block endogenous peroxidase and with 0.05 N NaOH in 40% ethanol to denature DNA. Mouse anti-O6-medG antibody (provided by Dr. Christopher P. Wild, University of Leeds) was used as the primary antibody. The specificity of this monoclonal antibody has been demonstrated previously (18). Rabbit antimouse IgG (Jackson, West Grove, PA) was the secondary antibody. The entire antibody-antigen complex was made visible by diaminobenzidine tetrahydrochloride (Sigma Chemical Co., St. Louis, MO). Liver O6-medG DNA adducts in AOM-injected animals were used as positive controls (17). Omission of primary antibody and the 0 time point were used as negative controls. At least 20 crypt columns/ animal were randomly chosen for analysis according to criteria established previously (34). The staining intensity was assessed by cell position within the crypt. Images of colonic crypts were captured on a MICROSTAR IV, Reichert microscope networked to a Sony DXC-970 MD 3CCD camera and a Power Macintosh computer. Images were processed

755

756


Fig. 12. Summary of data describing DNA adduct level (—), DNA repair enzyme expression (z z z z), and apoptosis (z - z) throughout the time course of the study. To place all three values on a common scale, numbers were chosen for each (50, 10, and 20, respectively), and the proportion of those numbers at each particular time point is shown on the graphs. Apoptosis as a means of DNA-damaged cell removal declines at the time when repair enzyme level starts to increase but well before the peak in DNA adduct level is reached.

using NIH Image software, version 1.61. Optimum offset and gain were determined by preanalysis of multiple darkly- and lightly-stained tissues to maximize the distribution of stain intensity so that small differences in staining were quantifiable. For accurate and consistent results, once established, the settings remained constant for all images. Background staining intensity was determined on 10 randomly obtained images per animal and subtracted from the staining intensity of target cells. In Vivo Measurement of Apoptosis. Apoptosis was measured on the basis of the specific binding of TdT-mediated dUTP-biotin nick end labeling (Oncor, Gaithersburg, MD). Paraformaldehyde-fixed, paraffin-embedded tissue sections were deparaffinized and rehydrated. Endogenous peroxidase was quenched by immersing the section in 0.3% H2O2 in methanol. Nuclei of tissue sections were stripped of proteins by incubation with 10 mg/ml proteinase K (Ambion, Austin, TX) and washed in double-distilled water. The 13 equilibrium buffer was applied to the tissue sections for ,30 min. The working solution (TdT-digoxigenin-11 dUTP solution with a ratio of 1:8) was applied and incubated in a prewarmed humid chamber at 37°C with agitation every 10 min. Subsequently, antidigoxigenin peroxidase was applied to the tissue section in a humid chamber. The entire antibody-antigen complex was visualized using diaminobenzidine tetrahydrochloride and counterstained in 1% methyl green. Positive control tissue sections were prepared by nicking DNA with DNase I (Ambion). In negative control tissue sections, deionized water was substituted for TdT in the working solution. Apoptotic cells were identified on the basis of a combination of positive staining and morphological criteria as described by Kerr et al. (35). More than 20 well-oriented crypts were used for these analyses. Crypt height in number of cells and the number and position of apoptotic cells were recorded. Crypt height in number of cells was divided by three (dividing the crypt into three compartments: bottom, middle, and top). The number of apoptotic cells in each compartment was recorded. The apoptotic index was determined by dividing the number of apoptotic cells by the total number of cells in the crypt column and multiplying by 100. The apoptotic index by location within the crypt was calculated by dividing the number of apoptotic cells in each third of the crypt by the total number of cells in that compartment and multiplying by 100. In Vivo Measurement of O6-Alkylguanine-DNA Alkyltransferase. Paraformaldehyde-fixed, paraffin-embedded tissue sections were deparaffinized and rehydrated. Endogenous peroxidase was quenched by immersing the sections in 3% H2O2 in methanol. Antigen accessibility was enhanced by microwave treatment in a 0.1 M sodium citrate solution (pH

6.0). To block nonspecific binding, tissue sections were incubated with avidin, biotin, and TNB buffer [0.1 M Tris-HCl (pH 7.5), 0.15 M NaCl, and 0.5% blocking reagent]. Rabbit antirat alkyltransferase (provided by Dr. Rhoderick H. Elder, University of Manchester, Manchester, United Kingdom) was used as the primary antibody. The specificity of this antibody has been described previously (19). Also, a strong correlation between quantitative immunohistochemistry and biochemical alkyltransferase activity (P , 0.0001) has been reported previously (24). Biotinylated goat antirabbit IgG was the secondary antibody. The antigen-antibody complex was visualized using the tyramide signal amplification system (NEN Life Science Products, Boston, MA). Omission of primary antibody was used as a negative control. At least 20 crypt columns/animal were randomly chosen for analysis. The staining intensity was assessed by cell position within the crypt as described above for measurement of DNA adducts. Epithelial cells in the left crypt column of at least 20 well-oriented crypts/animal were captured, and the staining intensity was plotted for each cell within the crypt. Background staining intensity was determined on 10 randomly obtained images/animal and subtracted from the staining intensity of target cells. Statistical Analysis. Data were analyzed to determine the effect of carcinogen administration using a one-way ANOVA. All responses by relative cell position in the colonic crypt were analyzed using PROC MIXED (random effects: rat and crypt; fixed effect: time after carcinogen exposure) of SAS (36). Comparisons across time (Figs. 1, 5, and 8) were made using the Student-Newman-Keuls multiple range test (37). To accommodate the intrasubject dependency, comparisons among the different crypt compartments (bottom, middle, and top; Figs. 4, 7, and 11) were performed by paired t test (37). To extract the most information from the graphs, in Figs. 12 and 13 we have expressed DNA adduct level, DNA repair enzyme, and apoptotic index as a fraction of conveniently chosen numbers (50, 10, and 20, respectively). This places all three values on a common scale and allows interpretation of percentage changes in time or in cell position. Significance was set at P , 0.05.

Acknowledgments We acknowledge the important contribution of Dr. Laurie A. Davidson, who optimized several of the immunohistochemical assays for this study; Dr. John Roths, who introduced us to quantitative immunohistochemistry; and Rong Cui and Dr. Yi-Hai Jiang, who helped with sample collection. We also acknowledge Dr. James Allan for critical reading of the manuscript.


Fig. 13. Summary of data describing DNA adduct level (—), DNA repair enzyme expression (z z z z), and apoptosis (- z -) at each time point, showing expression by relative cell position. To place all three values on a common scale, numbers were chosen for each (50, 10, and 20, respectively), and the proportion of those numbers at each location within the crypt is shown on the graphs. DNA adduct level is distributed relatively uniformly, with a slight but significantly higher level in the bottom portion of the crypt at all time points. The DNA repair enzyme level is significantly higher in the upper portion of the crypt until 12 h, when repair enzyme levels in the bottom one-third of the crypt increase significantly. Levels of apoptosis are always higher in the bottom one-third of the crypt than in the other two segments.

References 1. Hawks, A., and Magee, P. N. The alkylation of nucleic acids of rat and mouse in vivo by carcinogen 1,2-dimethylhydrazine. Br. J. Cancer, 30: 440 – 446, 1974.

9. Jacoby, R. F., Lior, X., Teng, B. B., Davidson, N. O., and Brasitus, T. A. Mutations in the K-ras oncogene induced by 1,2-dimethylhydrazine in preneoplastic and neoplastic rat colonic mucosa. J. Clin. Investig., 87: 624 – 630, 1991.

2. James, J. T., and Autrup, H. Methylated DNA adducts in the large intestine of ICR/Ha and C57BL/Ha mice given 1,2-dimethylhydrazine. J. Natl. Cancer Inst., 70: 541–546, 1983.

10. Hollstein, M., Sidransky, D., Vogelstein, B., and Harris, C. C. p53 mutations in human cancers. Science (Washington DC), 253: 49 –53, 1991.

3. Saffhill, R., Margison, G. P., and O’Connor, P. J. Mechanisms of carcinogenesis induced by alkylating agents. Biochim. Biophys. Acta, 823: 111–145, 1985.

11. Montesano, R., Hall, J., Hollstein, M., Mironov, N., and Wild, C. P. Alkylation repair in human tissues. In: B. M. Sutherland and A. D. Woodhead (eds.), DNA Damage and Repair in Human Tissues, pp. 437– 452. New York: Plenum Publishing Corp., 1990.

4. Hall, C. N., Badawi, A. F., O’Connor, P. J., and Saffhill, R. The detection of alkylation damage in the DNA of human gastrointestinal tissues. Br. J. Cancer, 64: 59 – 63, 1991. 5. Rogers, K. J., and Pegg, A. E. Formation of O6-methylguanine by alkylation of rat liver colon, and kidney DNA following administration of 1,2-dimethylhydrazine. Cancer Res., 37: 4082– 4087, 1977. 6. Likhachev, A. J., Margison, G. P., and Ruggero, M. Alkylated purines in the DNA of various rat tissues after administration of 1,2-dimethylhydrazine. Chem. Biol. Interact., 18: 235–240, 1977. 7. Boldogh, I., Ramana, C. V., Chen, Z., Biswas, T., Hazra, T. K., and Grosch, S. Regulation of expression of the DNA repair gene O6-methylguanine-DNA methyltransferase via protein kinase C-mediated signaling. Cancer Res., 58: 3950 –3956, 1998. 8. Burmer, G. C., and Loeb, L. A. Mutations in the K-ras 2 oncogene during progressive stages of human colon carcinoma. Proc. Natl. Acad. Sci. USA, 86: 2403–2407, 1989.

12. Gerson, S. L., Allay, E., Vitantonio, K., and Dumenco, L. L. Determinants of O6-alkylguanine-DNA alkyltransferase activity in human colon cancer. Clin. Cancer Res., 1: 519 –525, 1995. 13. Sancar, A. DNA repair in humans. Annu. Rev. Genet., 29: 69 –105, 1995. 14. Thompson, C. B. Apoptosis in the pathogenesis and treatment of disease. Science (Washington DC), 267: 1456 –1462, 1995. 15. Hirose, Y., Yoshimi, N., Makita, H., Hara, A., Tanaka, T., and Hideki, M. Early alterations of apoptosis and cell proliferation in azoxymethaneinitiated rat colonic epithelium. Jpn. Cancer Res., 87: 575–582, 1996. 16. Meikrantz, W., Bergom, M. A., Memisoglu, A., and Samson, L. O6Alkylguanine DNA lesions trigger apoptosis. Carcinogenesis (Lond.), 19: 369 –372, 1998. 17. Asamoto, M., Mikheev, A. M., Ziang, Y. Z., Wild, C. P., Hall, J., and Montesano, R. Immunohistochemical detection of DNA alkylation adducts

757

758


in rat and hamster liver after treatment with dimethylnitrosamine. Exp. Pathol., 41: 71–78, 1991. 18. Wild, C. P., Smart, G., Saffhill, R., and Boyle, J. M. Radioimmunoassay of O6-methyldeoxyguanosine in DNA of cells alkylated in vitro and in vivo. Carcinogenesis (Lond.), 4: 1605–1609, 1983. 19. Chinnasamy, N., Rafferty, J. A., Margison, G. P., O’Connor, P. J., and Elder, R. H. Induction of O6-alkylguanine-DNA-alkyltransferase in hepatocytes of rats following treatment with 2-acetylaminofluorene. DNA Cell Biol., 16: 493–500, 1997. 20. Potten, C. S., Li, Y. Q., O’Connor, P. J., and Winton, D. J. A possible explanation for the differential cancer incidence in the intestine, based on distribution of the cytotoxic effects of carcinogens in the murine large bowel. Carcinogenesis (Lond.), 13: 2305–2312, 1992. 21. Rao, C. V., Nayini, J., and Reddy, B. S. Effect of oltipraz [5-(2pyrazinyl)-4-methyl-1,2-dithiol-3-thione] on azoxymethane-induced biochemical changes related to early colon carcinogenesis in male F344 rats. Proc. Soc. Exp. Biol. Med., 197: 77– 84, 1991. 22. Grisham, J. W., Greenberg, D. S., Kaufman, D. G., and Smith, G. J. Cycle-related toxicity and transformation in 10T1/2 cells treated with N-methyl-N9-nitro-N-nitrosoguanidine. Proc. Natl. Acad. Sci. USA, 77: 4813– 4817, 1980. 23. Smith, G. J., Kaufman, D. G., and Grisham, J. W. Decreased excision of O6-methylguanine and N7-methylguanine during the S phase in 10T1/2 cells. Biochem. Biophys. Res. Commun., 92: 787–794, 1980. 24. Zaidi, N. H., Liu, L., and Gerson, S. L. Quantitative immunohistochemical estimates of O6-alkylguanine-DNA alkyltransferase expression in normal and malignant human colon. Clin. Cancer Res., 2: 577–584, 1996. 25. Celik, C., Lewis, D. A., and Mittleman, A. Induction of colon mucosal b-glucuronidase production as a mechanism for 1,2-dimethylhydrazine colon carcinogenesis. J. Surg. Oncol., 24: 209 –211, 1983. 26. Ijiri, K., and Potten, C. S. Response of intestinal cells of differing topographical and hierarchical status to ten cytotoxic drugs and five sources of radiation. Br. J. Cancer, 47: 175–185, 1983. 27. Ijiri, K., and Potten, C. S. Further studies on the response of intestinal crypt cells of different hierarchical status to eighteen different cytotoxic agents. Br. J. Cancer, 55: 113–123, 1987.

28. Preston-Martin, S., Pike, M. C., Ross, R. K., Jones, P. A., and Henderson, B. E. Increased cell division as a cause of human cancer. Cancer Res., 50: 7415–7421, 1990. 29. Engelward, B. P., Allan, J. M., Dreslin, A. J., Kelly, J. D., Wu, M. M., Gold, B., and Samson, L. D. A chemical and genetic approach together define the biological consequences of 3-methyladenine lesions in the mammalian genome. J. Biol. Chem., 273: 5412–5418, 1998. 30. Toft, N. J., Winton, D. J., Kelly, J., Howard, L. A., Dekker, M., te Riele, H., Arends, M. J., Wyllie, A. H., Margison, G. P., and Clarke, A. R. MSH2 status modulates both apoptosis and mutation frequency in the murine small intestine. Proc. Natl. Acad. Sci. USA, 96: 3911–3915, 1999. 31. Chang, W-C. L., Chapkin, R. S., and Lupton, J. R. Predictive value of proliferation, differentiation and apoptosis as intermediate markers for colon tumorigenesis. Carcinogenesis (Lond.), 18: 721–730, 1997. 32. Ruifrok, A. C. C., Weil, M. M., Thames, H. D., and Mason, K. A. Diurnal variations in the expression of radiation-induced apoptosis. Radiat. Res., 149: 360 –365, 1998. 33. Van Benthem, J., Vermeulen, E., Winterwerp, H. H. K., Wild, C. P., Scherer, E., and Den Engelse, L. Accumulation of O6-methylguanine and 7-methylguanine in DNA of N-nitroso-N-methylbenzylamine treated rats is restricted to non-target organs for N-nitroso-N-methylbenzylamine-induced carcinogenesis. Carcinogenesis (Lond.), 13: 2102–2105, 1992. 34. Lee, D. Y., Lupton, J. R., Aukema, H. M., and Chapkin, R. S. Dietary fat and fiber alter rat colonic mucosal lipid mediators and cell proliferation. J. Nutr., 123: 1808 –1817, 1993. 35. Kerr, J. F. K., Gobe, G. C., Winterford, C. M., and Harmon, B. V. Anatomical methods in cell death. Methods Cell Biol., 46: 1–27, 1995. 36. SAS Institute, Inc. SAS Technical Report P-229. Cary, NC: SAS Institute, Inc., 1992. 37. Ott, R. L. An Introduction to Statistical Methods and Data Analysis, pp. 291–295; 822– 824. Belmont, CA: Wadsworth, Inc., 1993.